Fingerstick potassium and creatinine strip with on-strip hemolysis rejection — the story so far

How to read this. This is a synthesis, not a new source of truth: every claim it cites carries the tag the candidate file gives it right now, copied not paraphrased, and every line quoted from a deck is quoted verbatim. See docs/narrative-spec.md.

The one-paragraph version

This is a single-use test strip and handheld meter that reads potassium and creatinine from one fingerstick capillary drop, with a third on-strip channel whose only job is to detect a hemolysed sample and suppress the potassium result rather than report a falsely elevated one — aimed at clinic staff drawing samples during outpatient heart-failure and CKD visits, to support in-visit RAAS-inhibitor dose titration instead of a venous draw sent out to a lab. The regulatory shape is real but asymmetric: both generic types are confirmed Class II [claim 1: verified] [claim 2: verified], and the creatinine channel has a genuine cleared predicate [claim 5: verified], but the potassium channel — the entire point of the device — has no predicate in this exact form factor, though the file's own audit found that framing overstated against the candidate's own interest [claim 7: unconfirmed]. Reimbursement, TAM, and FTO remain blank rather than weak: no CMS data file exists locally to check payment, coverage, or volume, and no patent search has ever run. What actually decides the candidate is unreached by any of that scoring: whether capillary sampling can be reliably screened for hemolysis at fingerstick volume before the result reaches a screen [claim 22: unconfirmed]. A business case was added to the file on 2026-09-03, and it layers twelve further unverified assumptions on top of everything above — who buys the device, whether the strip clears margin at the assumed price, and whether the calendar and capital to reach first commercial dollar are survivable — none of them checked by any Verifier pass [claim 44: unverified].

The story so far

RAASi therapy improves outcomes in heart failure and CKD but raises potassium, so dose titration is gated on a lab result that today comes back after the patient has left the clinic — the clinical premise behind this candidate, stated from a secondary source and unchecked at Stage 3 for both the drug-discontinuation rates it cites [claim 20: unconfirmed] and the monitoring cadence it assumes makes an in-visit result valuable [claim 21: unconfirmed]. The domain dossier this candidate links carries a more directly sourced version of the comorbidity side of that premise: a 418-site US heart-failure registry found CKD present in 64% of hospitalizations by discharge eGFR, a different and better-sourced number than anything in this candidate's own claims, though it measures a sicker, hospitalized population rather than the outpatient titration population this device targets [dossier: potassium-monitoring-ckd].

The mechanism is a genuine engineering bet, not a paperwork problem. Capillary lancing lyses red cells, and the resulting free potassium falsely elevates the result — "the named reason a strip-based fingerstick potassium meter plausibly does not already exist," in the file's own words [claim 22: unconfirmed]. The proposed fix suppresses the affected result rather than correcting it, which trades a wrong answer for no answer, so the real performance metric is the invalid-result rate in ordinary clinic hands, not accuracy on clean samples [claim 23: unconfirmed]. No connector in this repo reaches that question; it is settled only by a paired capillary-versus-venous bench study. The domain dossier's own abandonment record for capillary sampling as a practice (not this product) is a real, sourced twin: a 2021 inpatient study found capillary specimen rejection "as high as 30% in the initial 30 patients enrolled" and repeat lancing needed in 15% of cases, and a 2025 four-center study put capillary hemolysis at roughly 1.7x venous [dossier: potassium-monitoring-ckd]. Read carefully in both directions, as the dossier itself insists: that people retreat from capillary potassium is evidence the failure mode is real and unsolved, which is exactly what this candidate proposes to fix — it is not evidence that the fix works, and it is not evidence anyone would adopt it if it did.

The regulatory picture is a genuine asymmetry, not a uniform gap. Both generic types are confirmed Class II — 21 CFR 862.1600 for potassium, 862.1225 for creatinine [claim 1: verified] [claim 2: verified] — with product codes CEM and CGL sitting under them [claim 3: verified] [claim 4: verified]. The creatinine channel has a real predicate anchor, K070068, cleared for capillary, venous and arterial whole blood at the point of care [claim 5: verified]. The potassium channel, the device's entire reason to exist, does not have one in this form factor — but the frontmatter's "no specific predicate identified" is, per the same Stage 3 audit, overstated against the candidate's own interest: a complete enumeration of both product codes under 862.1600 over thirty years found two cleared whole-blood, point-of-care potassium systems (K200865, K042270), just none of them strip-and-fingerstick [claim 7: unconfirmed]. Whether FDA would even route this device through a Dual 510(k) and CLIA Waiver pathway at all is structurally unconfirmable here, since no connector reads FDA guidance [claim 8: unconfirmed].

Reimbursement and TAM carry the same shape: not weak evidence, no evidence. The favorable half of the reimbursement story — that CPT 84132 already says "whole blood" so no new code is needed — runs entirely through AMA-licensed descriptor text this pipeline can never resolve [claim 11: unconfirmed] [claim 12: unconfirmed], and the deciding question, whether the CLFS allowable clears the strip's cost, has no figure on either side [claim 14: unconfirmed]. No market number is asserted anywhere in the file; the only checkable-in-principle TAM component is blocked by a missing CMS cache and would in any case measure mostly central-lab draws this device cannot take [claim 16: unconfirmed]. FTO sits in the same place: zero patents have been searched, PATENTSVIEW_API_KEY is unset, and the named likely estates are the same firms already practising in this exact art [claim 25: unconfirmed] [claim 27: unconfirmed]. Layered on top of all of it are thirteen unverified claims the three decks surfaced and wrote back into the file — from whether prescribers act on a suppressed result [claim 31: unverified] to whether the strip's calibration holds across the anemic hematocrit range these patients present with [claim 41: unverified].

The file's updated: field moved from 2026-08-29 to 2026-09-03 because a business case was added, and what it added is not new evidence but a franker accounting of the commercial assumptions the earlier scoring left implicit. It records, each as its own unverified claim never checked by anyone: that the outpatient clinic itself, not a distributor, GPO, or health-system central lab, is the direct purchaser [claim 44: unverified]; that a net per-test strip price can be set below the combined CLFS allowable for 84132 and 82565 [claim 45: unverified] while the manufacturer still earns positive gross margin on the strip's bill of materials [claim 46: unverified]; that the handheld meter itself is not a barrier to adoption at whatever price or placement it carries [claim 47: unverified]; that a single clinic's insourced volume is enough to make the economics work [claim 48: unverified], against a material share of clinics choosing this device over an already-cleared incumbent analyzer [claim 49: unverified]; that the calendar [claim 50: unverified] and capital [claim 51: unverified] to reach first commercial dollar, and the lag before a clinic's first paid CLFS claim [claim 52: unverified], are all within a financeable range; that the hemolysis-detection-and-suppression module could be qualified as an add-on to an already-cleared point-of-care analyzer without triggering a fresh full 510(k) on the host device [claim 53: unverified]; that a lay-comprehension, untrained-operator human-factors package for an eventual OTC submission is affordable on top of the professional-track submission [claim 54: unverified]; and that a US health-system-level study could ever affordably and usefully establish a measured link between monitoring frequency and RAASi/MRA discontinuation or its downstream cost [claim 55: unverified]. Every one of these twelve claims is recorded, in the file's own words, "so the Verifier can see it" — none of them has actually been checked.

What's solid

• 21 CFR 862.1600 exists, is titled "Potassium test system," and classifies it Class II — read verbatim from the eCFR, including that its identification names serum, plasma and urine and does not name whole blood, capillary or fingerstick, a genuine gap in the predicate argument that is carried rather than waved off [claim 1: verified]
• 21 CFR 862.1225 exists, is the "Creatinine test system" regulation, Class II — read verbatim, narrower still than 862.1600 (plasma and urine only, not even serum) [claim 2: verified]
• Product code CEM ("Electrode, Ion Specific, Potassium") sits under 862.1600 at Class II [claim 3: verified]
• Product code CGL ("Electrode, Ion Based, Enzymatic, Creatinine") sits under 862.1225 at Class II and is the product code on K070068 [claim 4: verified]
• K070068 is a real FDA 510(k) — Nova StatSensor Creatinine Hospital Meter, cleared 2007, Substantially Equivalent — with an Indications for Use covering capillary, venous and arterial whole blood at the point of care, calibrated to give "plasma equivalent results," the documented mechanism by which a whole-blood device sits inside a plasma-worded generic type; its scope is the creatinine channel only [claim 5: verified]
• K091547 is a real FDA 510(k) — the Nova Max Plus glucose/β-ketone meter, a fingertip capillary-drop strip cleared for home lay use — not a potassium or creatinine predicate, but a real format analog behind the OTC/home-use route this candidate carries as a future submission rather than a launch claim [claim 6: verified]

What's still open

• Whether the potassium channel has an adequate predicate: no cleared strip-and-fingerstick potassium device exists, but cleared whole-blood point-of-care potassium systems do, and which of them counts as a predicate is an FDA substantial-equivalence judgment no database answers [claim 7: unconfirmed]
• Whether FDA's Dual 510(k) and CLIA Waiver by Application pathway is available as described — structurally unconfirmable, since no connector reads FDA guidance documents [claim 8: unconfirmed]
• Whether the Abbott i-STAT CHEM8+ cartridge's potassium channel is cleared and CLIA moderate complexity — the only K-number located for it turned out to be an ionized-calcium clearance [claim 9: unconfirmed]
• Whether a CLIA-waived potassium test already exists on the Piccolo Xpress platform — the underlying clearance for whole-blood point-of-care potassium is verified separately, but the waived-list half was never checked [claim 10: unconfirmed]
• Whether CPT 84132 and 82565 exist with descriptors that fit this device — AMA-licensed text this pipeline cannot resolve by design [claim 11: unconfirmed] [claim 12: unconfirmed]
• Whether a waived-certificate site would need to append the QW modifier [claim 13: unconfirmed]
• The load-bearing negative economic claim: whether the CLFS allowable for a single electrolyte clears the strip's cost to insource it — no dollar figure exists on either side [claim 14: unconfirmed]
• Whether any Medicare NCD or LCD restricts point-of-care potassium testing — an unexamined gap, not an absence of restrictions [claim 15: unconfirmed]
• All three decomposed TAM components — Medicare Part B volume [claim 16: unconfirmed], net price per strip [claim 17: unconfirmed], and the reachable-share assumption of 10-20% of switching clinics within five years [claim 18: unconfirmed] — plus the aggregator-grade HF/CKD prevalence figures the file itself says must never be multiplied by anything [claim 19: unconfirmed]
• The clinical-premise figures — RAASi/MRA cessation rates after a hyperkalemia diagnosis [claim 20: unconfirmed] and the guideline-derived monitoring cadence [claim 21: unconfirmed] — both taken from a search summary rather than a primary source or the guideline itself
• The mechanism claim that decides the candidate: whether capillary hemolysis falsely elevates potassium reliably enough that on-strip detection can suppress it, settled only by bench data no connector can reach [claim 22: unconfirmed], and its consequence — that the real performance metric is the invalid-result rate in ordinary hands, not clean-sample accuracy [claim 23: unconfirmed]
• The domain dossier's abandonment record for capillary sampling as a practice: a 2021 inpatient study reported specimen rejection "as high as 30% in the initial 30 patients enrolled" and repeat lancing in 15% of cases, and a 2025 four-center study found capillary hemolysis at roughly 1.7x venous — evidence the underlying sampling problem is real, not evidence this candidate's suppression mitigation works [dossier: potassium-monitoring-ckd]
• The asserted moat (hemolysis detect-and-suppress logic, membrane chemistry, strip yield curve) [claim 24: unconfirmed] and the FTO field beneath it — zero patents searched, no number named, and the named likely estates already active in this exact art [claim 25: unconfirmed] [claim 27: unconfirmed]
• Capital intensity and time-to-revenue, recorded as the file's own honest principal weakness, with the 510(k)-versus-De Novo fork entirely unsearched [claim 26: unconfirmed]
• The three stakeholder assumptions with no connector able to reach them: clinicians acting on an in-visit result rather than deferring to a confirmatory draw [claim 28: unconfirmed], clinics obtaining and absorbing a CLIA Certificate of Waiver [claim 29: unconfirmed], and patients accepting an extra fingerstick [claim 30: unconfirmed]
• Thirteen assumptions the three decks surfaced and wrote back as unverified, never yet checked by anyone: whether a suppressed result changes what a prescriber does [claim 31: unverified], whether staff can raise an adequate drop from edematous or frequently-lanced patients [claim 32: unverified], whether the fingerstick replaces the venous draw rather than adding to it [claim 33: unverified], whether prescribers require published outpatient capillary-versus-venous evidence first [claim 34: unverified], whether a count of addressable clinic sites is obtainable at all [claim 35: unverified], whether payment exceeds the fully-loaded cost of running the test in office [claim 36: unverified], whether the clinic itself captures the payment for an insourced test [claim 37: unverified], whether the two analytes are separately payable [claim 38: unverified], whether an FDA-recognized analytical performance criterion exists to validate against [claim 39: unverified], whether the hemoglobin channel can detect free hemoglobin at the bias-relevant threshold at fingerstick volume [claim 40: unverified], whether plasma-equivalent calibration holds across the anemic hematocrit range these patients present with [claim 41: unverified], whether the strip holds its calibration and hemolysis threshold across a commercial shelf life [claim 42: unverified], and whether FDA would require the suppression function to be evaluated within the marketing application itself [claim 43: unverified]
• Twelve further assumptions a 2026-09-03 business case appended, none yet checked: that the outpatient clinic itself, not a distributor, GPO, or health-system central lab, is the direct purchaser [claim 44: unverified]; that a net strip price can be set below the combined CLFS allowable for 84132 and 82565 [claim 45: unverified] while the manufacturer still earns positive gross margin [claim 46: unverified]; that the meter's price or placement is not itself a barrier to adoption [claim 47: unverified]; that a single clinic generates enough insourced volume to make the economics work [claim 48: unverified], against a material share of clinics choosing this device over an already-cleared incumbent analyzer [claim 49: unverified]; that the calendar [claim 50: unverified] and capital [claim 51: unverified] to first commercial dollar, and the lag to a clinic's first paid CLFS claim [claim 52: unverified], are all financeable; that the hemolysis-suppression module could be qualified as an add-on to an already-cleared analyzer without a new full 510(k) [claim 53: unverified]; that a lay-use human-factors package is affordable alongside the professional submission [claim 54: unverified]; and that a health-system-level study could affordably establish a measured link between monitoring frequency and RAASi/MRA discontinuation or its downstream cost [claim 55: unverified]

Where the three lenses agree — and where they don't

All three decks were built 2026-08-25 and still check out clean (check_decks.py) against the candidate file as it stands now — the business case added on 2026-09-03 appended new claims at the end of the file without moving any claim number or tag a deck already cited, so the tags inside the quotes below are current rather than historical.

The desirability deck turns on whether anyone in the referral chain will actually act on the number. The file records this in its own words — that the value proposition dies while the device works if it fails [claim 28: unconfirmed] — and the deck names it the single most consequential thing to check before any capital is committed, ahead of the engineering and the margin questions the other two lenses turn on.

"whether an outpatient heart-failure or nephrology prescriber will actually change a RAASi or MRA dose within the visit on a fingerstick potassium result, instead of ordering a confirmatory venous draw and deferring" — the load-bearing condition of the desirability deck (built 2026-08-25).

The viability deck turns on a margin inequality with no verified figure on either side. Both the CLFS allowable [claim 14: unconfirmed] and the fully-loaded strip cost to the clinic [claim 36: unverified] are unquantified, and the deck names this the cheapest unresolved question in the file — one public CMS download and one internal cost estimate, answerable before any prescriber is interviewed or any patent is searched. The business case's own margin assumptions [claim 45: unverified] [claim 46: unverified] sit on top of this same unresolved inequality rather than settling it.

"whether the CLFS allowable for potassium and creatinine exceeds the fully-loaded per-test cost of running the strip in an outpatient clinic" — the load-bearing condition of the viability deck (built 2026-08-25).

The feasibility deck turns on the piece of physics everything else assumes will be solved: not whether the strip is accurate on a clean sample, but how often it produces no usable result at all in real clinic hands, because the mitigation suppresses rather than corrects.

"the invalid-result rate in ordinary clinic hands — what fraction of fingerstick attempts, by real clinic staff on real heart-failure and CKD patients, end with the potassium result suppressed rather than reported" — the load-bearing condition of the feasibility deck (built 2026-08-25).

The load-bearing question

If the on-strip hemoglobin channel can be shown to reliably suppress a hemolysed draw without voiding an unworkable share of ordinary fingersticks, and the CLFS allowable turns out to clear the strip's fully-loaded cost with room for the manufacturer's own margin, does an outpatient heart-failure or nephrology prescriber actually change a RAASi or MRA dose on that number in the visit — and does the clinic itself, rather than a distributor or health-system lab, capture enough of that insourced volume, at a price and a calendar the business can finance — or does an in-visit result, however accurate and however priced, just add a stick to a workflow that still ends in a confirmatory venous draw?